Toshiba AI Technology Catalog

  • Prediction candidate presentation
  • Operation and Control
  • Operation plan
  • Sensor data recognition
  • Status estimation
  • Media recognition

Predicting other vehicles’ behavior

Predicts future risks, to enable saver, smoother automatic driving.

  • Predicts the position and trajectory of the other vehicle a few seconds ahead, based on the shape of the intersection and lane information.
  • Detects locations where pedestrians might suddenly appear from a blind spot, using only the vehicle’s onboard sensors.


  • Strategic decision making in automatic driving based on risk prediction
  • Automatic brakes and other Advanced Driver-Assistance Systems (ADAS)
  • Vehicle-to-infrastructure (V2I) communications; e.g., notification of vehicles that present potential risks
  • Route planning for transport robots that do not interfere with workers in logistics warehouses
  • Other cases of mobile robot control requiring attention to moving objects nearby

Benchmarks, strengths, and track record

  • Can be applied to intersections with any shapes, to enable simultaneous predictions of which lane will be used when multiple lanes are available.
  • Predicts behavior on any place without preparing high-precision maps in advance. 
  • Predicts the shape of roads and obstacles using multiple candidates, where blind spots prevent narrowing down to a single candidate.
  • Accepted by top international conferences on robots and image processing.
  • International Conference on Robotics and Automation (ICRA) 2020
  • Intelligent Transportation Systems Conference (ITSC) 2019
  • Intelligent Vehicles Symposium (IV) 2018
  • Winter Conference on Applications of Computer Vision (WACV) 2021


Please include the title “Toshiba AI Technology Catalog: Predicting other vehicles’ behavior” or the URL in the inquiry text.
Please note that because this technology is currently the subject of R&D activities, immediate responses to inquiries may not be possible.


  • A. Kawasaki and A. Seki, "Multimodal Trajectory Predictions for Urban Environments Using Geometric Relationship between a Vehicle and Lanes,” ICRA, 2020.
  • T. Sugiura and T. Watanabe, “Probable Multi-hypothesis Blind Spot Estimation for Driving Risk Prediction,” ITSC, 2019.
  • A. Kawasaki and T. Tasaki, “Trajectory Prediction of Turning Vehicles based on Intersection Geometry and Observed Velocities,” IV, 2018.
  • A. Kawasaki and A. Seki, "Multimodal Trajectory Predictions for Autonomous Driving Without a Detailed Prior Map,” WACV, 2021.